Projects
Disaster Risk Prediction Platform
Python, Scikit-learn, Docker, Kubernetes, REST API
Architected and developed a full-stack disaster risk prediction platform, implementing a Random Forest Classifier in Python using Scikit-learn to predict event likelihood.
Engineered a data ingestion pipeline to process and clean 5 GB of geospatial and meteorological data from public APIs, performing feature engineering to improve model performance.
Deployed the application as a containerized service on Docker also using Kubernetes, providing a REST API for front-end visualization.
ChronoVault – Mental Wellness Platform
Python, MongoDB, GitHub Actions, AI
Designed and built a full-stack web application using a Python backend and MongoDB database to provide users with mental wellness tools and ethically-grounded AI engagement features.
Developed a content recommendation engine by performing exploratory data analysis (EDA) on user interaction data to identify engagement patterns.
Managed the complete software development lifecycle, including CI/CD pipelines with GitHub Actions for automated testing and deployment.
Privacy-First AI Personal Assistant
Developer & ArchitectPython, Google Agent Development Kit (ADK), Gemini API, Vertex AI
Developed a "Human-in-the-Loop" (HITL) AI agent that prioritizes data privacy by requiring explicit user consent before storing sensitive long-term memories.
Architected a custom Long-Running Operation (LRO) workflow to pause agent execution upon detecting sensitive keywords (e.g., health data) and resume only after receiving user validation.
Implemented sophisticated context engineering and event compaction to ensure rejected data is immediately purged from short-term session history.